MRC (1+3) PhD Studentship in Sheffield: (Bayesian Decision Analysis in 
Clinical Research Programmes)

Objective: To develop a career path in statistical and health economic 
modelling research.

Place:  A Medical Research Council Capacity Building Studentship is 
currently available in the Centre for Bayesian Statistics and Health 
Economics (CHEBS - http://www.shef.ac.uk/chebs/) which is the 
collaboration between Health Economics & Decision Science section at the 
School of Health and Related Research (http://www.shef.ac.uk/heds/), and 
the Dept of Probability and Statistics (http://www.sheffield.ac.uk/pas/) 
at the University of Sheffield.

Title: Bayesian Decision Analysis in Clinical Research Programmes – 
adding costs and benefits to decision rules and design

Stipend: approximately £12,940 per annum. The scholarship offers a 1+3 
year option, with the 1st year providing an opportunity to gain a Master 
in Public Health (Health Services Research).

Who can apply?
The successful candidate will need a 1st or 2:1 or MSc level training in 
a numerate subject/discipline such as Statistics, Operational Research, 
Mathematics, Economics, Industrial Engineering, Management Science, 
Physics or Systems Control.

Who to contact:
For further enquiries on the research / career opportunity this 
represents contact Dr Alan Brennan via email: [log in to unmask]
or to apply with a CV send to Ms Charlotte Hollins via email: 
[log in to unmask]


Deadline for applications:
20 June 2008

Research Aims:
The research question concerns how best to incorporate Bayesian decision 
analysis to clinical research programmes. Our interest is in health 
economic analysis at an earlier stage of decision making concerning 
clinical trials.

Background:
The project will review of existing Bayesian and health economic 
approaches to study analysis and design.  The theory of 
cost-effectiveness analysis of new interventions via evidence synthesis 
and health economic modelling are relatively well developed.  Similarly, 
standard frequentist approaches to specifying study design and analysis 
are implemented hundreds of times per year across the world. 
Nevertheless it is a constant refrain that the trial evidence does not 
provide enough evidence for decision makers and that modelled extensions 
of the evidence are required e.g. from interim (biomarkers) to final 
outcomes (clinical events), short to long-term survival, clinical or 
disease specific outcome to generic quality of life etc.

Case Studies:
(1) The first proposed collaboration is on the ESTEEM trial, examining 
long term outcomes of hormonal therapies versus surgery in women over 
75.  This is a long-term (up to 9 years) trial which has recently begun. 
  There is some existing prior evidence on these treatments but not in 
elderly women.  The issue is essentially one of competing risks – women 
incur loss of utility due to surgery in the short-term, NHS costs are 
higher too and the women may die of other illnesses anyway before any 
benefit on survival is derived.
(2) The second case study involves a suite of studies known as the IBIS 
II trials on anastrazole in post-menopausal women at increased risk of 
breast cancer.  This involves likely additional benefit in reduced 
breast cancer incidence which will be weighed against the possibility of 
increased osteoporosis risks.  ScHARR is world leading in the economic 
modelling of osteoporosis, and has recently undertaken NICE appraisals 
and guidelines work.

The case studies were chosen in conjunction with Prof Rob Coleman and Dr 
Lynda Wyld of the North Trent Cancer Research Network who are 
investigators on the trials and Prof. Julia Brown (Dir of Clinical 
Trials Research Unit, University of Leeds) part of the National Cancer 
research network, all of whom are keen to collaborate on this project

Research tasks:
In each case study, the researcher will first build a quantified 
decision model of the key relevant interventions, disease pathways and 
resulting costs and health outcomes including the link to quality 
adjusted life years (QALYs).  This will involve searching for and 
synthesising evidence from published and other sources.  This will then 
be used to undertake probabilistic sensitivity analysis (PSA) and 
expected value of perfect information (EVPI) calculations to identify 
where key uncertainties exist.  Linked to this will be the development 
of a Bayesian Clinical Trial Simulation model for the clinical trial of 
interest.  This will be used to simulate the possible results from the 
proposed design.  It will also be used to simulate alternative proposed 
designs which might have been / be chosen.  In each case the simulated 
BCTS results will be returned to the decision model to see how the trial 
results would affect decisions on the treatments (e.g. assuming a NICE 
appraisal given the simulated evidence).  Trials designs which provide 
an increased chance of a clear rather than inconclusive decision are 
more valuable. The expected value of sample information (EVSI) process 
can quantify this as well as the remaining uncertainty for decision 
makers.  The researcher will also compute costs for alternative trial 
designs and undertake a quantified economic trade-off, computing the 
expected value of the trial versus the expected cost of obtaining it.

Value Added Summary:
The decision modelling and analysis of clinical trial designs from this 
economic perspective will shed light on whether proposed designs are 
optimal in terms of value to policy decision makers.  It is unlikely, 
but possible in theory, that the analysis could show that trials are 
unnecessary if enough existing prior information is available. More 
likely it will show that alternative sample sizes, follow-up periods, 
comparator arms, or other data collection such as resource use, quality 
of life data or epidemiology on incidence of recurrence, metastases etc 
would be optimal.

First Supervisor: Dr Alan Brennan (Health Economic Modelling)
http://www.shef.ac.uk/scharr/sections/heds/staff/brennan_a.html

Second Supervisor: Dr Jeremy Oakley (Bayesian Statistics)
http://jeremy-oakley.staff.shef.ac.uk/

Eligibility:
The full MRC scholarship is open to UK nationals, to EU nationals who 
work in the UK and to EU nationals who have been living in the UK for a 
period of 3 years. A partial scholarship (i.e., fees only) may be 
available to those living outside the UK but within the EU.  Those 
living outside of the EU are not eligible.  Anyone who is interested in 
applying is strongly recommended to view the full details regarding 
eligibility on the MRC website first:
http://www.mrc.ac.uk/Utilities/Documentrecord/index.htm?d=MRC002630

-- 
Dr Alan Brennan (BSc, ARCS, MSc, PhD)
Director of Health Economics and Decision Science,
ScHARR, School of Health and Related Research, University of Sheffield

*Our new proposed MSc in Health Economics and Decision Modelling*
www.shef.ac.uk/scharr/prospective_students/masters/hedsmsc
My Website: www.shef.ac.uk/scharr/sections/heds/staff/brennan_a.html
Section website: www.shef.ac.uk/scharr/sections/heds
Address: Regent Ct, 30 Regent St, Sheffield S1 4DA, ENGLAND
Tel:+44 (0)114 2220684   Fax:+44 (0)114 2724095